—
Ray RLlib
by Anyscale
A production-grade distributed reinforcement learning library supporting multi-agent training, offline learning, and external simulators, compatible with PyTorch and TensorFlow.
Reinforcement Learning
Agents learn optimal behavior by interacting with an environment and maximizing reward.
by Anyscale
A production-grade distributed reinforcement learning library supporting multi-agent training, offline learning, and external simulators, compatible with PyTorch and TensorFlow.
by DLR-RM
Open-source implementations of deep reinforcement learning algorithms in PyTorch with consistent interfaces and documentation.
by Google DeepMind
A free physics engine for simulating articulated structures and contact dynamics, widely used for robotics research, optimal control, and reinforcement learning.
by OpenAI
A policy gradient reinforcement learning algorithm that balances sample complexity and simplicity through clipped probability ratios.
by Google DeepMind
A foundational deep reinforcement learning algorithm combining Q-learning with neural networks, using experience replay and target networks.